Net dollar retention rate measures how recurring revenue from the same customer cohort changed over a period after expansion, contraction, and churn.
It is commonly called net revenue retention, or NRR. NDRR and NRR usually refer to the same cohort concept, but a company should define the name, revenue basis, and period in its metric policy.
The NDRR formula
Freeze the customers present at the beginning of the period.
NDRR = (starting recurring revenue + expansion - contraction - churn) ÷ starting recurring revenue x 100
Exclude revenue from customers acquired after the opening date. Otherwise the metric mixes retention with acquisition.
Worked example
An opening cohort contributes $500,000 in recurring revenue.
- expansion: $45,000;
- contraction: $20,000;
- churn: $35,000.
The ending cohort revenue is:
$500,000 + $45,000 - $20,000 - $35,000 = $490,000
NDRR is:
$490,000 ÷ $500,000 x 100 = 98%
The company replaced most lost recurring revenue with expansion, but the opening cohort still ended smaller.
NDRR versus gross revenue retention
Gross revenue retention removes expansion:
GRR = (starting recurring revenue - contraction - churn) ÷ starting recurring revenue x 100
GRR cannot exceed 100%. NDRR can.
Read both because expansion can hide a weak retention floor. A small set of large upsells may keep NDRR above 100% while many customers downgrade or leave. Stripe's 2026 guidance makes the same distinction: NRR includes expansion; GRR focuses on revenue held without it.
The cohort rules
Document:
- opening and closing dates;
- monthly, annual, or another recurring revenue basis;
- treatment of usage-based and variable revenue;
- currency conversion;
- mergers, migrations, and account-parent changes;
- pauses, credits, and reactivations;
- acquired or divested customers;
- what qualifies as expansion, contraction, and churn.
Keep the opening cohort fixed. If an account changes name or billing entity, preserve the identity mapping so operational housekeeping does not appear as churn and new business.
Segment before diagnosing
An aggregate NDRR may combine different systems. Break it down by:
- product and plan;
- customer size;
- industry and use case;
- geography;
- contract age;
- implementation path;
- acquisition source;
- account owner;
- high-touch versus self-serve service model.
Then show the movement bridge:
| Movement | Revenue | Accounts | Primary reason |
|---|---|---|---|
| Opening cohort | |||
| Expansion | |||
| Contraction | |||
| Churn | |||
| Ending cohort |
Revenue and account counts answer different questions. One enterprise churn can dominate dollars; many small-logo losses can expose a product or onboarding problem.
Connect retention to the revenue system
NDRR is a lagging outcome. Diagnose it through earlier evidence:
- promised use case and qualification;
- time to first value;
- adoption by user and feature;
- support or service intensity;
- executive sponsorship;
- renewal-risk signals;
- pricing and packaging changes;
- expansion trigger and buyer;
- product reliability;
- customer outcome.
The original ICP matters. If poor-fit customers churn after expensive implementation, acquisition may be the first problem to repair. Use the ICP development workshop to connect customer economics back to account selection.
Common mistakes
- adding new-logo revenue to the numerator;
- changing the cohort during the period;
- mixing MRR and ARR;
- reporting NDRR without GRR;
- combining recurring and one-time service revenue;
- treating missing account mappings as churn;
- comparing periods with different billing or currency policies;
- celebrating expansion without inspecting logo and gross-dollar loss.
NDRR is powerful because it compresses the customer base's revenue movement into one comparable ratio. It becomes useful only when the cohort is frozen, the movement bridge is retained, and operators can trace the outcome back to customer fit, adoption, value, and service evidence.